Yu Hao Blasts Xiaohongshu's Algorithm: Who's Right in the Platform vs. Creator Debate?

A Public Showdown Over Algorithmic Power

Recently, Yu Hao publicly took aim at Xiaohongshu on social media, targeting the platform's recommendation algorithm, content evaluation system, and the massive base of anonymous users — commonly known as the "momos." His fiery remarks quickly ignited public debate, with supporters and critics on both sides, launching a broader discussion about the boundaries of algorithmic power on platforms.

Yu Hao's core criticisms centered on three levels: First, is Xiaohongshu's recommendation algorithm "holding creators hostage," forcing them to cater to traffic-driven logic rather than authentic expression? Second, does the platform's evaluation mechanism create a distorted ecosystem where "data determines merit"? Third, are the masses of anonymous users using the default "momo" nickname destroying the quality of community discourse?

The Algorithm Black Box: Creators' Collective Anxiety

Yu Hao's criticism of Xiaohongshu's algorithm is far from an isolated case. In fact, this dissatisfaction has long been widespread among creators. Xiaohongshu's recommendation system is built on deep learning models that leverage user profiles, content tags, engagement data, and other multidimensional signals for personalized content distribution. While this AI-driven recommendation mechanism improves user experience and platform efficiency, it also creates a massive "algorithm black box."

Creators widely report that identical-quality content posted at different times can yield vastly different traffic results. Even more baffling is that the platform has never disclosed transparent distribution rules, leaving creators to guess algorithmic preferences through a near-mystical process of trial and error. This asymmetric power structure of information keeps creators perpetually in a passive position.

From a technical perspective, recommendation algorithms are optimized for maximizing overall platform user retention and engagement metrics, not for maximizing individual creators' interests. This means the algorithm fundamentally serves the platform's commercial objectives, with creators being just one variable in this optimization equation. Yu Hao's frustration, in a sense, strikes at the very heart of this structural contradiction.

The "Momos": The Double-Edged Sword of Anonymity

Yu Hao's criticism of the "momos" has also resonated widely. On Xiaohongshu, a large number of users operate under the platform's default-assigned "momo" nickname and avatar, participating in interactions in a nearly fully anonymous manner. These anonymous users tend to be far more uninhibited in comment sections, with the proportion of negative reviews and malicious attacks significantly higher than that of identified users.

Behind this phenomenon lies a clear behavioral psychology logic — anonymity reduces perceived social accountability, making users more prone to releasing aggression. However, from the platform's perspective, the "momos" also contribute vast amounts of engagement data, and engagement data happens to be the core fuel for recommendation algorithms. In other words, the algorithm to some extent "rewards" content that sparks controversy and emotional reactions, with anonymous users being active participants in this feedback loop.

This creates a fascinating closed loop: the algorithm recommends controversial content → anonymous users flood the comments → engagement data surges → the algorithm further increases recommendation weight. Creators are caught in the middle, needing traffic to survive while enduring the psychological toll of anonymous attacks.

Xiaohongshu's Position: The Platform Governance Dilemma

From Xiaohongshu's perspective, things are far from simple. As a content platform with over 300 million monthly active users, Xiaohongshu must balance multiple stakeholder interests: users' content consumption experience, creators' creative incentives, advertisers' commercial demands, and the overall health of the community atmosphere.

On the anonymous user issue, Xiaohongshu faces a classic dilemma. Mandating real-name identification could significantly lower user participation thresholds and engagement activity, while allowing unchecked anonymity could continue to deteriorate the community ecosystem. Currently, Xiaohongshu has been attempting to curb malicious behavior through technical measures such as AI content moderation and comment filtering, but the results have clearly not yet met creators' expectations.

Regarding algorithm transparency, content platforms worldwide face similar challenges. Fully disclosing algorithm rules could lead to large-scale engagement manipulation and "SEO-style" content gaming, which would further degrade the content ecosystem. This is a classic paradox in technical governance, not a problem unique to Xiaohongshu.

The Deeper Question: Platform Responsibility in the Age of AI Algorithms

The dispute between Yu Hao and Xiaohongshu may appear to be a conflict between an individual and a platform, but it actually reflects a much larger question in the age of AI algorithms — when recommendation algorithms become the core infrastructure for information distribution, what kind of public responsibility should platforms bear?

The EU's Digital Services Act already requires large platforms to provide basic explanations of their recommendation algorithms and allow users to choose non-personalized content sorting options. In China, the Cyberspace Administration previously issued the "Provisions on the Management of Algorithmic Recommendations for Internet Information Services," explicitly requiring algorithm recommendation service providers to "offer users options that are not based on their personal characteristics."

From this regulatory trend, Yu Hao's demands are not without merit. Platforms do need to find a better balance between algorithmic efficiency and transparency, and they need to provide creators with clearer rule expectations.

However, it should also be noted that attributing all problems in a platform's ecosystem to algorithms is an oversimplification. The health of a content ecosystem depends on the combined effects of platform rules, user literacy, creator mindset, business models, and many other factors. The algorithm is just one — albeit a very important — component.

Looking Ahead: Dialogue Is More Valuable Than Confrontation

The positive significance of this debate is that it has once again pushed the important issue of "the relationship between algorithms and creators" to the center of public attention. Regardless of who is ultimately more "right," this kind of open discussion itself is driving the platform ecosystem toward a healthier direction.

For Xiaohongshu, earnestly listening to creator feedback and making tangible improvements in algorithm transparency and anonymity governance is essential for maintaining the platform's long-term competitiveness. For creators, understanding the complexity of platform operations and developing a rational understanding of algorithmic mechanisms while reasonably expressing their demands is equally indispensable.

In an era where AI algorithms increasingly dominate content distribution, what platforms and creators need is not a zero-sum game, but a symbiotic relationship built on transparent rules. This debate may well be an opportunity to foster exactly that kind of relationship.